INSIGHTS INTO PERSONAL ICT USE: … · Self-tracking, life-logging, quantified-self, personal...

WI-

736

University of Augsburg, D-86135 Augsburg Visitors: Universitätsstr. 12, 86159 Augsburg Phone: +49 821 598-4801 (Fax: -4899) University of Bayreuth, D-95440 Bayreuth Visitors: Wittelsbacherring 10, 95444 Bayreuth Phone: +49 921 55-4710 (Fax: -844710) www.fim-rc.de

Discussion Paper

INSIGHTS INTO PERSONAL ICT USE: UNDERSTANDING CONTINUANCE AND DISCONTINUANCE

OF WEARABLE SELF-TRACKING DEVICES

by

Arne Buchwald1, Albert Letner2, Nils Urbach, Matthias von Entreß-Fürsteneck

1 At the time of writing this paper, Arne Buchwald was a research assistant at the Research Center Finance & Information Management and the Department of Information Systems Engineering & Financial Management at the University of Augsburg.

2 Federal Office for Migration and Refugees, Germany

presented at: Twenty-Sixth European Conference on Information Systems, Portsmouth, United Kingdom, June 2018

Twenty-Sixth European Conference on Information Systems (ECIS2018), Portsmouth,UK, 2018

INSIGHTS INTO PERSONAL ICT USE:

UNDERSTANDING CONTINUANCE AND DISCONTINU-

ANCE OF WEARABLE SELF-TRACKING DEVICES

Research paper

Arne Buchwald, EBS Business School, Germany ([email protected])

Albert Letner, Federal Office for Migration and Refugees, Germany

([email protected])

Nils Urbach, University of Bayreuth, Germany ([email protected])

Matthias von Entreß-Fürsteneck, University of Bayreuth, Germany

([email protected])

Abstract

Wearable self-tracking devices become increasingly common in our society and reflect the trend to-

wards the digitized individual. However, little is known what drives the continuance and discontinu-

ance usage of such devices. To empirically analyze factors leading to continuance and discontinuance

of wearable self-tracking device usage, we develop a conceptual model based on established post-

adoption concepts and the dual-factor theory. We validate it by applying structural equation modeling

on collected survey data among 357 self-tracking users. Whereas previous research focuses on pre-

dicting continuance intentions, our results reveal discontinuance to be another decisive determinant of

usage behavior. Additionally, we provide insights into factors driving continuance and discontinuance

intentions and derive practical implications for producers. Our results advance the theoretical dis-

course on IS post-adoption behavior in a personal ICT context.

Keywords: Self-tracking, quantified-self, wearable computing, information systems continuance, in-

formation systems discontinuance, technology use, personal ICT.

1 Introduction

Self-tracking, life-logging, quantified-self, personal analytics, and personal informatics are terms for

the current trend in personal information and communication technology (ICT) to collect and analyze

specific features of life on a regular basis through mobile and wearable digital devices (Lupton,

2014a). Self-tracking devices are placed in the category of wearable electronics and/or multi-sensor

platforms in the field of the Internet of Things (Swan, 2012). These devices can take the shape of

smartwatches, wristband sensors, wearable sensor patches, artificial reality-augmented glasses, brain

computer interfaces, or wearable body metric textiles (Swan, 2012). They enable the individual to cap-

ture daily activities, exercises, vital parameters, disease symptoms, or nutrition, among others (Gimpel

et al., 2013; Pantzar and Ruckenstein, 2015). Next to an increasing interest in self-tracking devices

from an academic point of view, we also observe a surge of interest in practice. Even though the pri-

vate collection and analysis of one’s personal data is not a completely new phenomenon (e.g., hand-

written data or with Excel charts), the topic becomes vibrant again due to new technologies, decreas-

ing sensor sizes, and increasing smartphone usage, all of which enable the user to do this practice in a

much more convenient way (Gimpel et al., 2013; Lupton, 2014b). Major players in the consumer elec-

tronic market, such as Apple, Google and Microsoft, as well as specialized producers, such as fitbit or

Buchwald et al. /(Dis-)Continuance of Self-Tracking Devices

Twenty-Sixth European Conference on Information Systems (ECIS2018), Portsmouth,UK, 2018 2

Jawbone, launched their own wearable self-tracking devices (e.g., Apple Watch, Android Wear, Mi-

crosoft Band, Fitbit Charge and Jawbone UP) and start to build up software and hardware ecosystems

around them. It is expected that the shipment of self-tracking devices will grow from 102 million units

in 2016 to more than 224 million units in 2020 (IDC, 2016).

The first generation of self-tracking devices (e.g. the Nike Fuelband) can be seen as products that gen-

erate revenues at the point of sale only and solely run tracking and analyzing software, which is pro-

vided by the device producer in an enclosed ecosystem. Hence, in terms of direct revenue generation

for the specific product through a particular customer, it is not important how long the customer uses

the device after the initial adoption (however, in terms of brand reputation and rebuy intention, an on-

going use is aspired). Additionally, due to the closed ecosystem, the collected self-tracking data is of

limited value since it provides insights for the self-tracking users but does not allow service enhance-

ments for users by new and traditional third-party providers.

In contrast, the second generation of self-tracking devices (e.g. the Apple Watch) resembles a platform

for an entirely new and open ecosystem of applications and services for new and traditional third-party

providers, which can create additional value beyond the pure tracking and analysis of data for the user

and revenue for themselves over time (e.g., personalized sport and fitness support, dynamic health-

insurances, and digital health-care support) (Lupton, 2014a). These applications and services have in

common that the associated business and service models rely on the continuous supply with data rec-

orded by the basic technology – the wearable self-tracking device. Consequently, for the success of

these applications and services, it is necessary that the self-tracking device is used regularly and con-

tinuously after the initial adoption.

Regular and ongoing usage can be predicted by continuance factors but also inhibited by discontinu-

ance factors. While such factors have been investigated in post-acceptance models (e.g. Bhattacherjee,

2001; Furneaux and Wade, 2011, 2017; Recker, 2014, 2016), these models do not fully account for the

domain of wearable self-tracking devices placed in the personal ICT context and were not synthesize

in one comprehensive dual-factor model that considers both types of important determinants of usage

behavior at the same time. Therefore, by identifying the determinants that promote and inhibit contin-

uous use behavior, we want to answer the following research question:

RQ: What determines continuance and discontinuance behavior of wearable self-tracking devices?

Concerning the adoption research stream, much research has focused on the initial acceptance of in-

formation systems (IS), especially in a business IS context. Also, in the post-acceptance phase useful

contributions on continuance use in a consumer context were published by Bhattacherjee (2001) and

Limayem et al. (2007), whereas Recker (2016) focused on discontinuance in a business IS context.

However, theoretical post-acceptance research with a focus on continuance and discontinuance inten-

tion in a personal ICT context is still scarce and will be the focus of this study:

Adoption phase Business IS context Personal ICT context

Initial acceptance Out of scope Out of scope

Post-acceptance Continuance Out of scope Focus of the study

Post-acceptance Discontinuance Out of scope

Table 1: Focus of the study

While previously developed models focus on either continuance or discontinuance, we aim at combin-

ing both facets in our study because they influence human behavior by initiating or preventing human

actions. Therefore, based on what we learn from existing theoretical approaches, we deductively de-

rive a conceptual model to answer our research question. To this end, we develop a set of 12 hypothe-

ses that we empirically evaluate using survey data collected from 357 self-tracking users and by apply-

ing structural equation modeling. Overall, we aim at advancing the theoretical understanding in the

field of personal ICT usage and particularly (dis)continuance research as well as at providing practical

implications for producers and third-party application developers in the self-tracking industry.



2 Theoretical Foundations

2.1 Wearable Self-Tracking Devices

Wearable self-tracking devices can be assigned to the category of personal ICT (together with

smartphones and tablet computers) since they are mobile (used on the users wrist), are adopted by in-

dividuals for their own personal usage and enable users to engage in various activities with one device

(Hong and Tam, 2006; Scheepers and Middleton, 2013). In the specific self-tracking domain, the liter-

ature basis is continuously growing. Gimpel et al. (2013) identified five motivational factors which

explain the individual’s activity of self-tracking, while Sjöklint et al. (2015) investigated how individ-

uals use self-tracking devices and how they cope with personal data provided in everyday life. Fur-

thermore, Pfeiffer et al. (2016) examined consumer acceptance factors of self-tracking devices, where-

as Baumgart and Wiewiorra (2016) analyzed what motivations to start self-tracking drive different

self-tracking activities and what role self-control plays on the self-tracking behavior. Regarding the

considered type of devices, technology acceptance research was conducted for wearable devices in

general (Spagnolli et al., 2014; Yang et al., 2016), smartwatches (Chuah et al., 2016; Kim and Shin,

2015) as well as fitness-trackers (Sol and Baras, 2016). These results in the context of self-tracking

and wearable devices serve as a promising basis for our research project.

2.2 Continuance and Discontinuance

Previous research extensively analyzes the adoption of IS, most often relying on the technology ac-

ceptance model (TAM) (Davis, 1989), the innovation diffusion theory (Moore and Benbasat, 1989) as

two prominent theories in this area. Subsequent research extending the original TAM by additional

factors converged in the unified theory of acceptance and use of technology (UTAUT) (Venkatesh et

al., 2003) and its extension (UTAUT2) (Venkatesh et al., 2012). The latter theories put forth additional

adoption determinants, such as social influence or hedonic motivation. While these acceptance models

made important contributions to the understanding of the initial technology adoption, they only pro-

vide limited insights into the post-acceptance phase (Sorgenfrei et al., 2014). For instance, Kim and

Malhotra (2005) showed that applying the untailored TAM does not capture the actual processes in-

volved in continued usage behavior or discontinuance. Furthermore, Karahanna et al. (1999) showed

how the behavioral intentions for IT adoption and usage are influenced very differently.

Looking at this discrepancy, Bhattacherjee (2001) opens up the domain of post-acceptance theories by

arguing based on the expectation confirmation theory that satisfied consumers will continue with IS

usage while dissatisfied consumers will discontinue. He zeroes in on continuance aspects and suggests

that usage is driven by the continuance intention which in turn is influenced by the satisfaction with

and the perceived usefulness of the IS. Several empirical studies were carried out to identify relevant

factors explaining usage. Petter et al. (2008) provide a summary of 21 studies in which use behavior

and other IS success-related variables are explained by only a few different factors on the individual

level, among them perceived usefulness. They show that no factor has strong literature support for ex-

plaining usage behavior. Therefore, only focusing on continuance aspects seems to be insufficient to

understand usage behavior.

To understand usage behavior more comprehensively, we also need to understand factors preventing

specific actions from taking place. By using theory of planned behavior or protection motivation theo-

ry as basis, previous research suggests that, similar to the continuance behavior, discontinuance deci-

sions are also based on formed intentions (Turel, 2016; Furneaux and Wade, 2017). Furneaux and

Wade (2011) are the first in the discontinuance domain, and they also focus on the business domain.

They reason that discontinuance is determined by (1) change forces such as environment or system

shortcomings as well as by (2) continuance inertia, e.g., system embeddedness or investments into it.

Although all of those factors are relevant for organizations, only some of them seem to be applicable

in a personal ICT discontinuance decision process. On the individual level, Recker (2014, 2016) con-

ceptualizes IS discontinuance in an individual’s working environment and suggests several factors



based on past works on technology use (perceived ease of use, perceived usefulness, perceived work

impediment) and status quo bias (perceived inertia and perceived sunk costs), all of which influence

the formation of the continuance and discontinuance intentions.

At this point, it is important to understand that continuance and discontinuance intentions are not con-

sidered to be one bipolar construct, but rather dual-factored constructs similar to the motivation-

hygiene theory (Herzberg, 1959, 1966) or the enabler inhibitor concept (Cenfetelli, 2004). They are

two independent parts with different determinants. For example, the occurrence of an error in a system

increases the intention to discontinue the use of a system. However, the absence of this error does not

necessarily build up the intention to continue using the system. Yet, both sides are relevant for under-

standing the reasons driving people’s usage behavior. Adoption models can benefit from a separate

view on both intention perspectives, by giving a deeper clarification on how well the continuance and

the discontinuance intentions can be explained by antecedent factors and on how important their re-

spective influence is on the actual usage. Current research on dual-factor theory has been conducted in

various contexts such as enablers and inhibitors of technology usage (Cenfetelli and Schwarz, 2011;

Recker, 2016), trust and distrust (Lewicki et al., 1998), and satisfaction and dissatisfaction (Herzberg,

1966).

To understand usage behavior in more detail, new research projects focusing on a particular research

context with contextual variables may also valuably extend generic theories with a more refined pic-

ture of determinants and their strengths (Alvesson and Kärreman, 2007; Johns, 2006; Venkatesh et al.,

2012). To illustrate this point in the case of personal ICT and self-tracking, devices collecting data on

people’s well-being, health and fitness may raise concerns, whether those data are shared with third

parties, e.g. an insurance company. Thus, trust placed in the manufacturer to not share information

may play an important role in the individual’s decision-making process to use a self-tracking device or

not. The idea of contextual factors suggests that variables explaining behavior differ from case to case

and offers an additional possible explanation as to why Petter et al. (2008) find no strong support in

the literature for any relationship explaining use behavior on the individual level.

Since previous research in a personal ICT context is scarce and not specifically carried out with the

aim to explain wearable self-tracking use behavior and particularly (dis)continuance in the post-

acceptance phase, we propose a comprehensive theoretical dual-factor model, including relevant fac-

tors from related literature, which we will explain subsequently.

3 Hypothesis Development

Having introduced previous research on continuance and discontinuance of IS usage behavior, we will

now explain the different constructs and proposed relationships to analyze use in the context of weara-

ble self-tracking devices. The context of wearable self-tracking devices refers to smartwatches (with

sophisticated self-tracking functions) and wristband sensors, such as fitness trackers. Furthermore, our

definition covers not only the hardware, but also the software and the associated ecosystem.

3.1 Use as well as Continuance and Discontinuance Intentions

IS continuance, IS continuance behavior, or IS continuous usage can be described as “behavioral pat-

terns reflecting continued use of a particular IS” (Limayem et al., 2007, p. 707). As a form of behav-

ior, it results from a consciously built intention and is not a one-time outcome but rather a behavior

that is caused by a cycle of repeating or changing situations in which an individual intendedly or au-

tomatically uses the IS until the individual decides to discontinue (Limayem et al., 2007). To intended-

ly using the system, the user builds up a continuance intention (CI), which basically reflects all factors

driving a person towards the conscious choice to continue the use. Analogous to the continuance inten-

tion, the discontinuance intention (DI) reflects all factors that lead to an individual’s conscious choice

to stop using an IS. For clarification purposes, however, we stress that discontinuance does not com-

prise the stopping of the activity (e.g., running) within which the self-tracking device was used, nor

discontinuance as a result of substitution for a similar product. In this first step, we particularly focus



on factors linked to the self-tracking device that may lead an individual to discontinue using the device

at all (while continuing the actual activity, e.g., running). Our proposed relationships are based on the

theory of planed behavior (Ajzen, 1985) and the theory of reasoned action (Fishbein and Ajzen, 1975)

in which several beliefs form intentions that ultimately result in behavior. Thus, we posit:

H1: Continuance intention is positively related to use.

H2: Discontinuance intention is negatively related to use.

3.2 Status Quo Bias

Continuance intentions are influenced by the status quo bias (Recker, 2014) which reflects “people’s

preference for maintaining their current status or situation” (Kim and Kankanhalli, 2009, p. 569). The

wish to remain in the status quo can result from an individual cost–benefit analysis and psychological

factors (Lee and Joshi, 2016). The first can be the outcome from people perceiving disadvantages of

change or discontinuance to be greater than the advantages (Samuelson and Zeckhauser, 1988). In the

context of self-tracking devices, discontinuance would generate little advantages, except that users

save the negligible effort of equipping and using it. However, since self-tracking devices collect data

over time, particularly a change or discontinuance would (usually) result in a loss of the accumulated

information and be perceived as a disadvantage. Additionally, users of self-tracking devices often find

themselves within an ecosystem or community around their device or its brand, sharing experience

and information. Stopping the use of the particular self-tracking device would also lead to losing the

connection to this ecosystem or community. Drawing on the concept of loss aversion as a psychologi-

cal theory in the area of human decision making (Kahneman and Tversky, 1979), even minor goal-

unrelated losses due to changing from the current situation to another could be perceived as larger than

they actually are and as a result, people try to avoid them. Since that losses are perceived larger than

gains in value perception, people prefer to remain in the status quo (Kim and Kankanhalli, 2009). This

leads to a will to continue the use and is referred to as cognitive-based inertia (PCBI) (Recker, 2014).

Thus, we posit:

H3: Perceived cognitive-based inertia has a positive effect on the continuance intention.

Considering different types of inertia, individuals can also form attachments to routines or systems by

affection, strengthening the individual’s status quo bias (Polites and Karahanna, 2012). This results

from the individual being comfortable and happy with the system or even when pleasure is taken in its

usage, leading to a positive emotional bond (Polites and Karahanna, 2012; Lin et al., 2015). They in-

troduce this type of inertia as affective-based inertia (PABI). Since individuals consciously pursue ac-

tions with a positive contribution to their lives, the affective-based inertia raises individuals’ continu-

ance intentions of the associated actions. In the case of self-tracking devices, the affective-based iner-

tia is formed during extensive every-day usage. Hence, we posit:

H4: Perceived affective-based inertia has a positive effect on the continuance intention.

Next to inertia, Recker (2014) identifies another economic phenomenon that leads an individual to

remain in the status quo, the evaluation of transition costs in comparison to sunk costs. The construct

perceived sunk costs (PSC) is defined as “comparison of transition costs such as time and effort of

adapting to a new situation versus the time and effort already invested in learning to use the existing

system” (Recker, 2014, p. 5). Transition costs are especially relevant in the case of switching to anoth-

er system or routine. However, within the context of self-tracking devices, neither continuance nor

discontinuance creates transition costs. Though, this does not apply to sunk costs, because after an in-

vestment is made, an individual wants to draw as much benefit as possible from the investment (due to

the money, time, and effort already invested), even when the individual can partially regain the finan-

cial investment by reselling the device. That consequently creates a status quo bias with the intention

to continue using the current system (Kim and Kankanhalli, 2009), regardless of whether reasons for

discontinuance are present (Polites and Karahanna, 2012). The extent to which an individual wants to

retain the status quo depends on the individual’s perception of the investment. Thus, we posit:

H5: Perceived sunk costs have a positive effect on the continuance intention.



3.3 Social Influence

Social influence in a consumer context is defined as “the extent to which consumers perceive that im-

portant others (e.g., family and friends) believe they should use a particular technology” (Venkatesh et

al., 2012, p. 159). While social influence has been referred to using the terms subjective norm, social

factors, and image in previous research (e.g., Ajzen, 1991; Moore and Benbasat, 1991; Thompson et

al., 1991), Venkatesh et al. (2003, p. 451) argue that “each of these constructs contains the explicit or

implicit notion that the individual’s behavior is influenced by the way in which others will view them

as a result of having used the technology”. Image is defined as “the degree to which use of an innova-

tion is perceived to enhance one’s image or status in one’s social system” (Moore and Benbasat, 1991,

p. 195). Consumer technology, such as mobile devices, may be regarded as a symbol for fashion and

wealth if the diffusion of the technology is not yet substantial (Sarker and Wells, 2003). Individuals

therefore adopt the technology to increase their self-importance (Lu et al., 2005). We argue that con-

sumer technology, such as self-tracking devices, are highly personal mobile devices, which are mostly

worn visibly daylong and therefore also function as a technology to increase one’s self-importance.

Furthermore, subjective norm refers to “the perceived social pressure to perform or not to perform the

behavior” (Ajzen, 1991, p. 188). Because self-tracking devices offer a new and unfamiliar way to col-

lect and analyze highly personal data, we argue that this practice may lead to a strong formation of

opinion and controversial discussion of one’s social group concerning the use.

While social influence is often discussed within the acceptance literature, it has not been considered so

far in the continuance and discontinuance literature. This may be because research shows that social

influence decreases with a growing experience with the technology (Venkatesh et al., 2003; Venkatesh

and Morris, 2000). In contrast, we argue that for vibrant and controversial new technologies, such as

self-tracking devices, social influence by one’s social group does not just occur within the acceptance

phase but continuously, and can change the intended effective direction due to new circumstances

(e.g., negative news about the device manufacturer). We further propose dividing social influence into

two separate variables – positive and negative social influence. It seems reasonable to argue that the

absence of a positive social influence is not naturally leading to an increased discontinuance intention.

However, we expect that a negative social influence of one’s social group affects the perception

whether to use a certain technology and therefore influences the discontinuance intention. Hence, we

define positive social influence (PSI) as the extent to which personal ICT users perceive that important

others believe they should continue to use a particular technology and negative social influence (NSI)

as the extent to which personal ICT users perceive that important others believe they should discontin-

ue to use a particular technology and posit:

H6: Positive social influence has a positive effect on the continuance intention.

H7: Negative social influence has a positive effect on the discontinuance intention.

3.4 System Characteristics

Following Recker’s (2016) line of reasoning, we assume that users distinguish between system charac-

teristic advantages and disadvantages of an IS. In terms of system characteristic advantages, the two

variables perceived usefulness (PU) and perceived ease of use (PEOU) are widely accepted within the

IS acceptance research stream. Initially used and defined by Davis (1985, 1989) for his technology

acceptance model (TAM) as ”the degree to which a person believes that using a particular system

would enhance his or her job performance” (PU) and ”the degree to which a person believes that using

a particular system would be free of effort” (PEOU), both variables have been used in extensions of

TAM and alternative IS acceptance models (Davis, 1989, p. 320) . While PU and PEOU are the main

antecedents in the acceptance stage, it seems plausible at first to argue that both variables also influ-

ence the continuance decision (Bhattacherjee, 2001).

While PU has been considered in several previous studies (Recker, 2014; Bhattacherjee, 2001; Li-

mayem et al., 2007; Recker, 2016), the integration of PEOU should be discussed in more detail. Refer-

ring to empirical studies, Bhattacherjee (2001) discards PEOU from his expectation-confirmation



model of IS continuance, following the line of reasoning that it becomes non-significant in the post-

acceptance stage. This is due to the fact that users gain experience with a system and resolve their

PEOU concerns. Similarly, Venkatesh et al. (2003, p. 450) within their Unified Theory of Acceptance

and Use of Technology (UTAUT) state that the effort expectancy, which is defined “as the degree of

ease associated with the use of the system”, becomes non-significant when the technology is used ex-

tensively and sustainably. We follow the line of reasoning of Bhattacherjee (2001) and Venkatesh et

al. (2003) and therefore do not include PEOU in our model. By transferring PU into the self-tracking

context, we redefine PU as the degree to which a person believes that using a self-tracking device

would enhance his or her personal living condition. We argue that data that is regularly and continu-

ously collected by a self-tracking device, shared with the associated ecosystem, and analyzed by the

user, can contribute positively to one’s well-being, fitness, and/or health. As long as this circumstance

is believed by a user, PU will have a positive effect on the continuance intention. Hence, we posit:

H8: Perceived usefulness has a positive effect on the continuance intention.

Concerning the system characteristic disadvantages, Recker (2016) includes in his model a variable

called perceived work impediment, which is defined as “the individual assessment of system use in

terms of a detriment to work task performance due to a need to comply with the ineffectual require-

ments of system use”. For a deeper understanding of which specific system characteristic disad-

vantages influence the discontinuance intention in a personal ICT context, we transfer perceived work

impediment into perceived routine constraints. In addition, we adapt two variables from Furneaux and

Wade (2011) which seem suitable in a personal ICT context – system unreliability and system capabil-

ity shortcomings – and finally include the variable trust, since in a holistic view of hardware, software

and ecosystem it also reflects a system characteristic (Mayer et al., 1995; Mayer and Davis, 1999).

System reliability is defined as “the extent to which a system can be counted on to perform its intend-

ed tasks” (Furneaux and Wade, 2011, p. 582) . They hypothesized that a reduced system reliability

leads to an increased replacement intention due to the fact that the continued use of an unreliable sys-

tem is seen as a risk to the business (Furneaux and Wade, 2011). We adapt this hypothesis to the per-

sonal ICT context of self-tracking, arguing that an unreliable system (SU) (e.g., unreliable measure-

ment of data) frustrates the user and therefore effects the discontinuance intention positively. We fur-

ther assume that it is thereby not important whether the unreliable data measurement is caused by a

software or a hardware defect. We posit:

H9: System unreliability has a positive effect on the discontinuance intention.

System capability shortcomings (SCS) is defined as “a limitation in the functionality of an IS that un-

dermines its ability to meet individual needs” (Furneaux and Wade, 2011, p. 582). In our case, it refers

to the perceived shortcomings of the self-tracking device and relate to the individual user’s needs in

terms of software, hardware, and the associated ecosystem. For instance, the self-tracking system may

not be compatible to the individual’s preferred smartphone application or may not be supported by the

smartphone’s operating system. Another shortcoming might be, when the collected data cannot be

shared with one’s social group due to an incompatibility with the ecosystem. We posit that capability

shortcomings of the self-tracking system ultimately contribute to an individual’s intention to discon-

tinue using the device. Thus, we posit:

H10: System capability shortcomings are positively related to the discontinuance intention.

Work impediment is defined as “the individual perception of system use as a detriment to work task

performance” (Recker, 2016, p. 48). We transfer the construct to the individual’s private domain, re-

name it to perceived routine constraints (PRC) and define it as the individual assessment of system use

in terms of a detriment to routine performance due to a need to comply with the ineffectual require-

ments of system use. While Recker (2016) referred to work impediments in general, we relate it to the

private domain and zero in on the perceived constraints on an individual’s daily routines due to the

compliance with the requirements of a self-tracking device. For instance, the self-tracking system may

disturb individual routines such as wearing specific clothes, such as suits. We posit:

H11: Perceived routine constraints are positively related to the discontinuance intention.



Trust (T) is defined as “the intention to accept vulnerability to a trustee based on positive expectations

of his or her actions” (Colquitt et al., 2007, p. 909) and is significantly influenced by the trustworthi-

ness of the trustee (Colquitt et al., 2007), which in this case is the company operating the IS. Perceived

trustworthiness includes the three dimensions ability, benevolence, and integrity of the trustee (Mayer

et al., 1995). Mayer and Davis (1999) posit that the relative importance of each of these three dimen-

sions changes depending on specific situations in which any of facets may be more salient compared

to the other facets. Ability refers to “that group of skills, competencies, and characteristics that allow a

party to have influence within some domain”, benevolence is “the extent to which a trustee is believed

to want to do good to the trustor, aside from an egocentric profit motive”, while integrity is “the

trustor's perception that the trustee adheres to a set of principles that the trustor finds acceptable”

(Mayer and Davis, 1999, p. 124). Related to self-tracking devices, the construct thus comprises the

ability of the self-tracking service provider to continuously protect the individual’s data, continuous

take actions in the individual’s best interest and integrative behavior. We suggest that trust is an im-

portant factor because the service provider continuously gathers and analyzes data from the individual

that is highly sensitive. If trust diminishes as judged by the individual user, we argue that this percep-

tion contributes to a discontinuance intention. We posit:

H12: Trust is negatively related to the discontinuance intention.

4 Research Method

We choose quantitative-empirical methods to validate the conceptual model because they allow for a

statistical generalization on the basis of results “representative of the whole population at a lower cost

than collecting the data for the whole population” (Saunders et al., 2009, p. 144). Subsequently, we

crafted a survey instrument. We began this process by using mostly existing measurement items as an

initial operationalization that we subjected to a card sorting process to improve the validity and relia-

bility of our final measurement instrument. Ultimately, we collected empirical data using our survey

instrument and analyzed it using structural equation modeling (Straub, 1989; Urbach and Ahlemann,

2010).

4.1 Construct Operationalization

We used established and validated measures for our measurement instrument wherever possible to

increase its validity (e.g. DeLone and McLean, 2003). We often adapted the items such that the self-

tracking device as the focus of our study is centrally reflected in each of the statements. Each of the

item statements was measured with a seven-point Likert scale (Likert, 1932) between (1 = I do not at

all agree; 7 = I do fully agree). All constructs are measured reflectively. To further enhance the con-

vergent and discriminant validity of our instrument, we decided to conduct a card sorting exercise

(Moore and Benbasat, 1991) using an online tool. A group of five IS researchers was asked to assign

the randomized items to their respective constructs. As a result of item assignments and additional

qualitative feedback, we decided to revise the phrasing of items to increase the clarity.

4.2 Data Collection

We collected data by administrating our survey instrument to current active users of wearable self-

tracking devices (e.g smartwatchtes or fitness trackers), which means that the users already own and

still use their device for the purpose of fitness, health or well-being tracking. Users who do not yet

own or already stopped using their devices were excluded from the survey because they do not have a

current perception about the determinants of continuance and discontinuance as actual users have and

would therefore distort the results. Further, personal and voluntary usage is assumed but not ensured,

since we did not account for users who were encouraged by third parties such as physicians or fitness

trainers. Respondents who indicated, subsequent to the introduction page of the online questionnaire,

that they are not active users of a self-tracking device, were excluded from participating in the online

questionnaire. To gather our data from respondents, we circulated the invitation message to participate



in our study in online social networks (e.g., weblogs, Facebook wall postings, Facebook groups, and

Twitter), online business networks (e.g., Xing and LinkedIn), and the e-learning system of the authors’

university. We decided in favor of openly circulating our invitation to allow for a snowball effect

within the social media. Overall, we received 680 responses. After excluding incomplete and non-

plausible answers (323 responses), we analyzed the remaining 357 responses. As we circulated the

invitation for participation anonymously, we cannot determine a response rate. We controlled for the

variables gender, age, income, country of origin, educational achievement, computer skills, used self-

tracking device type, self-tracking device usage time, technical enthusiasm and health consciousness.

The survey was filled out by 78 female and 272 male respondents. 7 preferred to not report on their

gender. The average age was 28 with an average monthly income between 3,000 and 4,000 US Dollars

after taxes. The average time of usage for the self-tracking devices was 18 months. 272 used a smart-

watch, 78 activity trackers, 3 another form and 4 gave no input. More than half of the sample group

sees itself as proficient in computer skills or better and is interested in trying out new technical devices

(technical enthusiasts). 66 contain a degree on the tertiary level (e.g. bachelor or master) while 276

contain a secondary degree (e.g. high school diploma). 3 stated to have no degree, 12 gave no input on

this matter. 283 agreed or strongly agreed regarding actively taking care of their health. 52 agreed to

some extent, 8 neither agreed nor disagreed, 7 disagreed to some extent, 2 disagreed and 5 gave no

information in this matter. We gathered 134 responses from the United States, 91 from Germany, 26

from Australia, 21 from United Kingdom, 12 from Canada, 7 from Netherlands, 5 from Belgium, and

45 from 28 other countries. 17 gave no input in this matter.

5 Data Analysis and Results

We tested the hypotheses and measurement properties with the structural equation modeling (SEM)

approach partial least squares (PLS) (Chin, 1998; Wold, 2004) using the software SmartPLS (Ringle et

al., 2015). We decided in favor of PLS for the data analysis because PLS has advantageous character-

istics in comparison to covariance-based approaches (e.g., comparatively complex research models,

high quantity of indicators and not very established measures) (Fornell and Bookstein, 1982; Chin and

Newsted, 1999). Adhering to the validation guidelines of Lewis et al. (2005), Straub et al. (2004), and

Urbach and Ahlemann (2010), we tested the reflective measurement model in terms of unidimension-

ality, internal consistency reliability, indicator reliability, convergent validity, and discriminant validi-

ty.

Unidimensionality describes how the measurement items of a latent construct’s property reflect only

the respective construct better compared to other variables (Gerbing and Anderson, 1988). We tested

for unidimensionality by conducting an exploratory factor analysis (EFA) using SPSS 24 and selected

the principle component analysis in combination with the Varimax rotation and the Kaiser normaliza-

tion. While the conceptual model includes 13 constructs, the EFA only suggested 12 factors on the

basis of the empirical data. Continuance intention and discontinuance intension both load on one fac-

tor which is not surprising because the items of both constructs are reversed. All other items only load

on their respective factor with coefficients of at least 0.607, thus exceeding the threshold of 0.600

which is deemed high (Gefen and Straub 2005).

We evaluated internal consistency reliability using Cronbach’s alpha (CA) and composite reliability

(CR). Both the values for CA and CR of all constructs exceed the threshold of 0.7 (Nunnally and

Bernstein, 1994) in Table 2, suggesting a high degree of internal consistency reliability.

Convergent validity assesses the extent to which an item positively correlates with other items of the

same construct. We evaluated the average variance extracted, suggested by Fornell and Larcker

(1981). As shown in Table 2, the AVE value of each of the constructs exceeds 0.5, indicating that the

variance caused by measurement errors are lower compared to the construct’s variances.



Cronbach's Alpha Composite Reliability Average Variance Extracted

Continuance intention 0.965 0.977 0.935

Discontinuance intention 0.971 0.981 0.945

Use 0.894 0.926 0.758

Negative social influence 0.958 0.967 0.854

Perceived affective based inertia 0.915 0.937 0.749

Perceived cognitive based inertia 0.939 0.960 0.890

Perceived routine constraints 0.877 0.907 0.662

Perceived sunk costs 0.801 0.881 0.711

Perceived usefulness 0.857 0.903 0.701

Positive social influence 0.959 0.973 0.924

System capability shortcomings 0.735 0.871 0.773

System unreliability 0.871 0.919 0.791

Trust 0.897 0.929 0.765

Table 2. Assessment of the Measurement Model

Indicator reliability shows to what degree a variable is consistent with its respective measure. A con-

struct’s reliability is determined independently from other constructs. We performed a confirmatory

factor analysis (CFA) using SmartPLS. With the exception of the third item of the system capability

shortcomings scale (SCS3) and the fifth item of positive social influence (PSI5) which we subsequent-

ly dropped, all loadings in our model exceed the threshold of 0.708, showing that the corresponding

constructs explain more than 50% of the variance of an item. We evaluated the significance of the in-

dicator loadings with the bootstrapping procedure with 10,000 resamples. All loadings were signifi-

cant at the 0.01 level.

Discriminant validity assesses the extent to which a construct truly differs from other constructs of the

conceptual model. To evaluate discriminant validity, we assessed the items’ cross-loadings and the

Fornell-Larcker criterion. Regarding the cross loadings, each item should show the highest loading on

the construct with which it is theoretically related in comparison to other constructs (Chin, 1998).

Therefore, we surmise that interchangeability of the indicators between the different constructs is not

given. To support this point further, we analyzed the difference between the square roots of the AVE

obtained for every single construct and the interconstruct correlation. This method demands the con-

structs to explain more of the variance in their own indicators compared to the variances they share

with other constructs (Fornell and Larcker, 1981). In the context of this study, the square root of the

AVE for every single construct is larger than their interconstruct correlations (see Table 3), which of-

fers further indication for sufficient dissimilarity of the constructs.

CI DI USE NSI PABI PCBI PRC PSC PU PSI SCS SU T

CI 0.967

DI -0.810 0.972

USE 0.488 -0.476 0.871

NSI -0.228 0.254 -0.092 0.924

PABI 0.621 -0.534 0.524 -0.166 0.865

PCBI 0.127 -0.097 0.061 0.017 0.146 0.943

PRC -0.349 0.343 -0.225 0.216 -0.259 0.053 0.814

PSC 0.164 -0.096 0.129 0.002 0.189 0.052 0.081 0.843

PU 0.688 -0.573 0.448 -0.220 0.576 0.012 -0.262 0.272 0.837

PSI 0.252 -0.195 0.133 0.055 0.282 0.078 0.039 0.175 0.341 0.961

SCS -0.230 0.247 -0.208 0.109 -0.313 0.024 0.211 0.016 -0.282 -0.159 0.879

SU -0.290 0.291 -0.283 0.116 -0.387 0.025 0.224 -0.029 -0.297 -0.048 0.533 0.889

T 0.338 -0.246 0.240 -0.221 0.390 0.140 -0.135 0.109 0.353 0.145 -0.162 -0.201 0.875

Table 3: Fornell-Larcker Criterion



Turning to our results, we found support for 9 out of our 12 hypotheses. Our model explains 55.5% of

the variance of the continuance intention, 21.7% of the variance of the discontinuance intention, and

25.7% of the variance of the usage. To set these values in perspective, research in IS usage typically

achieves around 30% explanatory value (Meister and Compeau, 2002). All beta coefficients, the corre-

sponding significant levels and R-squares are summarized in Table 4.

Independent variables Dependent variables Beta coefficients P-values R²

Continuance intention Use 0.298 0.001*** 0.257

Discontinuance intention Use -0.235 0.014*

Perceived affective based inertia Continuance intention 0.324 0.000***

0.555

Perceived cognitive based inertia Continuance intention 0.076 0.019*

Perceived sunk costs Continuance intention -0.038 0.240ns

Perceived usefulness Continuance intention 0.515 0.000***

Positive social influence Continuance intention -0.015 0.440ns

Negative social influence Discontinuance intention 0.146 0.006**

0.217

Perceived routine constraints Discontinuance intention 0.243 0.000***

System capability shortcomings Discontinuance intention 0.077 0.319ns

System unreliability Discontinuance intention 0.150 0.018*

Trust Discontinuance intention -0.138 0.010**

* significant at p < .050; ** significant at p < .010; *** significant at p < .001

Table 4: Results

6 Discussion

In line with existing theory (Limayem et al., 2007), our results show that the continuance intention

strongly predicts the use of a wearable self-tracking device. In addition, we also found evidence for

discontinuance intention as a significantly negative determinant of use in accordance with our concep-

tualization on the basis of the motivation-hygiene theory (Herzberg, 1959). Continuance intention re-

flects factors based on positive beliefs, which facilitate continued use behavior by increasing user sat-

isfaction (Thong et al., 2006; Sorgenfrei et al., 2014; Recker, 2016). In contrast, hygiene factors can

cause dissatisfaction – but not satisfaction –, conceptualized as discontinuance intention which reflects

factors based on negative beliefs (Recker 2016). For example, the presence of system unreliability fos-

ters a discontinuance intention, whereas its absence does not contribute to the formation of a continu-

ance intention. Accordingly, we contribute to the extension of the post-acceptance research stream by

showing that a dual-factor conceptualization of continuance and discontinuance intentions helps ex-

plaining the use of self-tracking devices in a personal ICT context. Further, from a practical point of

view, the significant negative influence of the discontinuance intention on use (β: -0.235) shows man-

ufacturers that they should not only consider factors fostering continuance intention but also consider

in their product development what factors lead to a user’s discontinuance intention and minimize

them.

Examining the status quo bias variables, our results only show strong support for the relationship be-

tween perceived affective based inertia and the continuance intention. As the influence of perceived

cognitive based inertia on the continuance intention is comparably weak (β: 0.076), we reject the hy-

pothesis. Additionally, the influence of sunk cost on the continuance intention is found to be not sig-

nificant. To explain the results partially deviating from our conceptualization, we suggest that affective

based inertia involves a positive emotional bond, and that a user wants to retain positive emotional

contributions to his or her life, which are generated by the use of a self-tracking device. Manufacturers

could therefore capitalize on this effect by including features, such as the personalization of the user

interface. In contrast, the loss of quantifiable but emotionless assets, such as collected information,

invested time, or money, seems to be negligible in this context. This explanation aligns with the find-



ings of Yu and Dean (2001) who investigated the role of emotional satisfaction in contrast to cognitive

satisfaction on customer loyalty. They showed that emotional satisfaction has a stronger impact on

customer loyalty than the cognitive satisfaction.

Concerning the social influence, we found support for the effect of negative social influence on the

discontinuance intention, but no support for positive social influence on the continuance intention. By

separating social influence into a positive and a negative dimension, we advance the current discourse.

The definition and operationalization of positive social influence is identical to the established con-

struct of social influence that has been proven to be relevant in the acceptance phase for wearable self-

tracking devices (Pfeiffer et al., 2016). A possible explanation for our divergent findings might be that,

in a post-adoption context, the potential loss of reputation within one’s social group that disagrees with

the use of a self-tracking device has a greater impact than the support of the use. In the case of positive

social influence, the social group of a user is merely supporting the initial decision to adopt a self-

tracking device, hence the user matches the expectation of his social group. On the other side, negative

social influence is in contrast with the initial decision of the user to use the self-tracking device, there-

fore it could force the user to reconsider his or her decision to use the device to meet his or her social

group expectations. With this in mind, producers of wearable self-tracking devices should invest in

ongoing marketing and service to avoid negative voices.

Furthermore, focusing on the system characteristics, we found support for perceived usefulness, sys-

tem unreliability, perceived routine constraints, and trust. The strong support for the influence of per-

ceived usefulness on continuance intention was confirmed in our data as hypothesized because the var-

iable is well established in this field of research and was proven to have a significant influence in vari-

ous contexts (e.g. Venkatesh et al., 2003; Venkatesh et al., 2012; Davis, 1989; Limayem et al., 2007;

Recker, 2016; Bhattacherjee, 2001; Davis, 1985). We also found weak support for the influence of the

system unreliability on the discontinuance intention. The results are in contrast to Furneaux and Wade

(2011) who tested the reverse variable system reliability in an organizational context and found no

support for their hypothesis. The different results could be explained with the distinct research con-

texts. We suggest that, while organizations often have IT-service departments and service contracts

with their vendors to solve reliability issues, within the personal ICT context it is nowadays expected

that a consumer technology is working reliable and accurate since users do often not have the

knowledge, time, or will for troubleshooting. Hence, it is important for producers of self-tracking de-

vices to update their devices regularly to prevent reliability problems.

Further, we found strong support for the influence of our newly developed variable perceived routine

constraints on the discontinuance intention. Considering the nature of a self-tracking device that is

worn and used almost permanently, our results show the need for these kinds of technology to inte-

grate seamlessly into the daily routines of the user to avoid the emergence of a discontinuance inten-

tion. Hence, during the development phase of the hardware, software and surrounding eco-system, the

focus should be in particular on the overall usability of the devices. Finally, trust has a negative impact

on the discontinuance intention, suggesting that users value a trustworthy vendor of a self-tracking

device, when their highly sensitive data is gathered and analyzed. While previous studies already con-

firmed in various contexts that trust into the vendor is an important factor (e.g. Gefen et al., 2003; Suh

and Han, 2002; Wang et al., 2003), we show that it is also important within the post-acceptance phase

in the domain of self-tracking devices and should therefore receive high attention by producers of

wearable self-tracking devices.

Finally, our results did not confirm the hypothesized influence of system capability shortcomings on

the discontinuance intention, in contradiction to the results of Furneaux and Wade (2011). While it

seems plausible that an information system is expected to fulfill its requirements continuously, we ex-

plain the diverging results with the research context of our study. In the personal ICT context, users

are nowadays able to anticipate the majority of potential shortcomings already during the acceptance

phase before the wearable self-tracking device is actually bought (e.g. with reviews or hands-on tests).

Hence, they are able to choose a device with zero or negligible shortcomings for themselves or are

willing to accept potential shortcomings. Consequently, producers of wearable self-tracking devices

should ensure that their products fulfill the major requirements already on release since their potential



customers already form a product opinion before the release of subsequent software and firmware up-

dates.

7 Conclusion

We set out to deductively build up and test a conceptual model with which we aim to explain an indi-

vidual’s continuous and discontinuance intention to use wearable self-tracking devices in a personal

ICT context. While research on the continuance and discontinuance of individual-level IS usage is yet

scarce, our study is one of the first that further explores this promising path and suggests a validated

and comprehensive dual-factor model in the personal ICT context. Our study makes two significant

contributions to the theoretical discourse. First, we show that hygiene factors (such as system unrelia-

bility, perceived routine constraints, trust and negative social influence by one’s social group) deter-

mine the conscious formation of a discontinuance intention. Second, our results also show that the

continuance intention is determined by the perceived usefulness and affective-based inertia.

Besides our promising results, we acknowledge the following limitations. Firstly, due to the chosen

distribution channels, our sample group consists mainly of self-tracking enthusiasts which is why we

cannot exclude a potential selection bias. A future broader validation of our model should ensure a

sample group consisting of more ordinary self-tracking device users. Secondly, in terms of the trade-

off between the width and the depth of a model, we decided in favor of a broad model because we

deem a validated broad model more valuable in the early stages of a research domain. Succeeding re-

search may then narrow down the focus on specific aspects and consider possible differences of

groups segmentations (e.g. self-tracking device type, gender, age, etc.).

Conclusively, focusing on the theoretical implications of our study, our proposed model is one of the

first to combine the research on continuance and discontinuance in a comprehensive model, therefore

building the basis for future research. Furthermore, by focusing our research on the field of self-

tracking, we transfer the current research of post-acceptance use into the personal ICT context. Con-

cerning the practical implications, producers of self-tracking devices and developers of third-party ap-

plications and services, especially with a focus on healthcare, well-being and fitness, get a deeper un-

derstanding which positive and negative factors concerning self-tracking devices are important for

customers and lead to a continuance or discontinuance intention. In detail, to generate a continuous

intention, it is not only important to develop hardware, software, and an associated ecosystem that cus-

tomers perceive to be useful but also enjoyable to use, all of which ultimately forms a positive emo-

tional bond. Possible measures could be the development of visually appealing devices with adaptable

styles (Pfeiffer et al., 2016) or gamification elements which challenge the user to sustain his or her

achievements or to reach new goals. In contrast, to inhibit the formation of a discontinuance intention,

self-tracking device developers should ensure that the device does not only provide the expected fea-

tures but also interacts seamlessly with the user. Hence, when designing the hardware and software,

factors such as used materials and visual appearance as well as software usability and interaction

should be considered to minimize the disturbance of the user’s daily routines as much as possible.



References

Ajzen, I. (1985). “From Intentions to Actions: A Theory of Planned Behavior.” In: Action Control. Ed. by J. Kuhl and J. Beckmann. Berlin, Heidelberg: Springer Berlin Heidelberg, p. 11–39.

Ajzen, I. (1991). “The Theory of Planned Behavior.” Organizational Behavior and Human Decision

Processes 50 (2), 179–211.

Alvesson, M. and D. Kärreman (2007). “Constructing mystery: Empirical matters in theory develop-

ment.” Academy of Management Review 32 (4), 1265–1281.

Baumgart, R. and L. Wiewiorra (2016)."The Role of Self-Control in Self-Tracking." In: Proceedings

of the Thirty Seventh International Conference on Information Systems. Dublin.

Bhattacherjee, A. (2001). “Understanding Information Systems Continuance: An Expectation-

Confirmation Model.” MIS Quarterly 25 (3), 351–370.

Cenfetelli, R. T. (2004). “Inhibitors and Enablers as Dual Factor Concepts in Technology Usage.”

Journal of the Association for Information Systems 5 (11), 472–492.

Cenfetelli, R. T. and A. Schwarz (2011). “Identifying and Testing the Inhibitors of Technology Usage

Intentions.” Information Systems Research 22 (4), 808–823.

Chin, W. W. (1998). “The Partial Least Squares Approach to Structural Equation Modeling.” In: Mod-

ern methods for business research. Ed. by G. A. Marcoulides. Mahwah, N.J: Lawrence Erlbaum,

p. 295–336.

Chin, W. W. and P. R. Newsted (1999). “Structural Equation Modeling Analysis with Small Samples

using Partial Least Squares.” In: Statistical strategies for small sample research. Ed. by R. H.

Hoyle. Thousand Oaks, Calif.: Sage Publications, p. 307–341.

Chuah, S. H.-W., P. A. Rauschnabel, N. Krey, B. Nguyen, T. Ramayah and S. Lade (2016). “Wearable

technologies: The role of usefulness and visibility in smartwatch adoption.” Computers in Human

Behavior 65, 276–284.

Colquitt, J. A., B. A. Scott and J. A. LePine (2007). “Trust, Trustworthiness, and Trust Propensity: A

Meta-analytic Test of their Unique Relationships with Risk Taking and Job Performance.” The

Journal of applied psychology 92 (4), 909–927.

Davis, F. D. (1985). A Technology Acceptance Model for Empirically Testing New End-User Infor-

mation Systems - Theory and Results.

Davis, F. D. (1989). “Perceived Usefulness, Perceived Ease of Use, and User Acceptance of Infor-

mation Technology.” MIS Quarterly 13 (3), 319–340.

DeLone, W. H. and E. R. McLean (2003). “The DeLone and McLean Model of Information Systems

Success: A Ten-Year Update.” Journal of Management Information Systems 19 (4), 9–30.

Fishbein, M. and I. Ajzen (1975). Belief, Attitude, Intention and Behavior: An Introduction to Theory

and Research. Reading, Mass.: Addidon Wesley.

Fornell, C. and F. L. Bookstein (1982). “Two Structural Equation Models: LISREL and PLS Applied

to Consumer Exit-Voice Theory.” Journal of Marketing Research 19 (4), 440–452.

Fornell, C. and D. F. Larcker (1981). “Evaluating Structural Equation Models with Unobservable Var-

iables and Measurement Error.” Journal of Marketing Research 18 (1), 39–50.

Furneaux, B. and M. Wade (2011). “An Exploration of Organizational Level Information Systems

Discontinuance Intentions.” MIS Quarterly 35 (3), 573–598.

Furneaux, B. and M. Wade (2017). “Impediments to Information Systems Replacement: A Calculus of

Discontinuance.” Journal of Management Information Systems 34 (3), 902–932.

Gefen, D., E. Karahanna and D. W. Straub (2003). “Trust and TAM in Online Shopping: An Integrat-

ed Model.” MIS Quarterly 27 (1), 51–90.

Gerbing, D. W. and J. C. Anderson (1988). “An Updated Paradigm for Scale Development Incorporat-

ing Unidimensionality and Its Assessment.” Journal of Marketing Research 25 (2), 186–192.

Gimpel, H., M. Nissen and R. Goerlitz (2013)."Quantifying the Quantified Self: A Study on the Moti-

vations of Patients to Track Their Own Health." In: Proceedings of the Thirty Fourth International

Conference on Information Systems. Milan.

Herzberg, F. (1959). The Motivation to Work. 2d ed. New York: Wiley.



Herzberg, F. (1966). Work and the nature of man. Oxford, England: World Publishing.

Hong, S.-J. and K. Y. Tam (2006). “Understanding the Adoption of Multipurpose Information Appli-

ances: The Case of Mobile Data Services.” Information Systems Research 17 (2), 162–179.

IDC (2016). “Worldwide Smartwatch Market Will See Modest Growth in 2016 Before Swelling to 50

Million Units in 2020.” URL:

https://www.businesswire.com/news/home/20160915005309/en/Worldwide-Smartwatch-Market-

Modest-Growth-2016-Swelling (visited on 04/19/2018).

Johns, G. (2006). “The Essential Impact of Context on Organizational Behavior.” Academy of Man-

agement Review 31 (2), 386–408.

Kahneman, D. and A. Tversky (1979). “Prospect Theory: An Analysis of Decision under Risk.” Econ-

ometrica 47 (2), 263–292.

Karahanna, E., D. W. Straub and N. L. Chervany (1999). “Information Technology Adoption Across

Time: A Cross-Sectional Comparison of Pre-Adoption and Post-Adoption Beliefs.” MIS Quarterly

23 (2), 183–213.

Kim, H.-W. and A. Kankanhalli (2009). “Investigating User Resitance to Information System Imple-

mentation: A Status Quo Bias Perspective.” MIS Quarterly 33 (3), 567–582.

Kim, K. J. and D.-H. Shin (2015). “An acceptance model for smart watches: Implications for the adop-

tion of future wearable technology.” Internet Research 25 (4), 527–541.

Kim, S. S. and N. K. Malhotra (2005). “A Longitudinal Model of Continued IS Use: An Integrative

View of Four Mechanisms Underlying Postadoption Phenomena.” Management Science 51 (5),

741–755.

Lee, K. and K. Joshi (2016). “Examining the use of status quo bias perspective in IS research: need for

re‐conceptualizing and incorporating biases.” Information Systems Journal.

Lewicki, R. J., D. J. McAllister and R. J. Bies (1998). “Trust and Distrust: New Relationships and Re-

alities.” Academy of Management Review 23 (3), 438–458.

Lewis, B. R., G. F. Templeton and T. A. Byrd (2005). “A Methodology for Construct Development in

MIS Research.” European Journal of Information Systems 14 (4), 388–400.

Likert, R. (1932). A Technique for the Measurement of Attitudes. New York: Columbia Univ.

Limayem, M., S. G. Hirt and C. M.K. Cheung (2007). “How Habit Limits the Predictive Power of In-

tention: The Case of Information Systems Continuance.” MIS Quarterly 31 (4), 705–737.

Lin, T.-C., S.-L. Huang and C.-J. Hsu (2015). “A Dual-factor Model of Loyalty to IT Product – The

Case of Smartphones.” International Journal of Information Management 35 (2), 215–228.

Lu, J., J. E. Yao and C.-S. Yu (2005). “Personal Innovativeness, Social Influences and Adoption of

Wireless Internet Services via Mobile Technology.” The Journal of Strategic Information Systems

14 (3), 245–268.

Lupton, D. (2014a)."Self-tracking Cultures: Towards a Sociology of Personal Informatics." In: Pro-

ceedings of the Australian Conference on Human-Computer Interaction (HCI). Sydney.

Lupton, D. (2014b)."Self-tracking Modes: Reflexive Self-Monitoring and Data Practices." In: Pro-

ceedings of the Imminent Citizenships: Personhood and Identity Politics in the Informatic work-

shop. Canberra.

Mayer, R. C. and J. H. Davis (1999). “The Effect of the Performance Appraisal System on Trust for

Management: A field Quasi-experiment.” Journal of Applied Psychology 84 (1), 123–136.

Mayer, R. C., J. H. Davis and F. D. Schoorman (1995). “An Integrative Model of Organizational

Trust.” Academy of Management Review 20 (3), 709–734.

Meister, D. B. and D. R. Compeau (2002)."Infusion of Innovation Adoption: An individual Perspec-

tive." In: Proceedings of the Annual Conference of the Administrative Sciences Association of

Canada 2002: Information Systems Stream. Winnipeg, Manitoba, Canada.

Moore, G. C. and I. Benbasat (1989). An Examination of the Implementation of Information Technolo-

gy for End Users: A Diffusion of Innovations perspective.

Moore, G. C. and I. Benbasat (1991). “Development of an Instrument to Measure the Perceptions of

Adopting an Information Technology Innovation.” Information Systems Research 2 (3), 192–222.

Nunnally, J. C. and I. H. Bernstein (1994). Psychometric Theory. 3. ed. New York: McGraw-Hill.



Pantzar, M. and M. Ruckenstein (2015). “The heart of everyday analytics: Emotional, material and

practical extensions in self-tracking market.” Consumption Markets & Culture 18 (1), 92–109.

Petter, S., W. DeLone and E. McLean (2008). “Measuring Information Systems Success: Models, Di-

mensions, Measures, and Interrelationships.” European Journal of Information Systems 17 (3),

236–263.

Pfeiffer, J., M. von Entress-Fürsteneck, N. Urbach and A. Buchwald (2016)."Quantify-Me: Consumer

Acceptance of Wearable Self-Tracking Devices." In: Proceedings of the Twenty-Fourth European

Conference on Information Systems (ECIS). Istanbul.

Polites, G. L. and E. Karahanna (2012). “Shackled to the Status Quo: The Inhibiting Effects of Incum-

bent System Habit, Switching Costs, and Inertia on New System Acceptance.” MIS Quarterly 36

(1), 21–42.

Recker, J. (2014)."Towards a Theory of Individual-level Discontinuance of Information Systems Use."

In: Proceedings of the Thirty Fifth International Conference on Information Systems. Auckland.

Recker, J. (2016). “Reasoning about Discontinuance of Information System Use.” Journal of Infor-

mation Technology Theory and Application 17 (1), 41–66.

Ringle, C. M., S. Wende and J.-M. Becker (2015). SmartPLS. Bönningstedt: SmartPLS GmbH.

Samuelson, W. and R. Zeckhauser (1988). “Status Quo Bias in Decision Making.” Journal of Risk and

Uncertainty (1), 7–59.

Sarker, S. and J. D. Wells (2003). “Understanding Mobile Handheld Device Use and Adoption.”

Communications of the ACM 46 (12), 35–40.

Saunders, M., P. Lewis and A. Thornhill (2009). Research Methods for Business Students. 5. ed. Har-

low: Financial Times Prentice Hall.

Scheepers, R. and C. Middleton (2013). “Personal ICT Ensembles and Ubiquitous Information Sys-

tems Environments: Key Issues and Research Implications.” Communication of the Association for

Information Systems 33.

Sjöklint, M., I. O. Constantiou and M. Trier (2015)."The Complexities of Self-tracking - An Inquiry

into User Reactions and Goal Attainment." In: Proceedings of the Twenty-Third European Confer-

ence on Information Systems. Münster.

Sol, R. and K. Baras (2016). “Assessment of activity trackers.” In: Proceedings of the 2016 ACM In-

ternational Joint Conference on Pervasive and Ubiquitous Computing Adjunct - UbiComp '16. Ed.

by P. Lukowicz, A. Krüger, A. Bulling, Y.-K. Lim and S. N. Patel. New York, New York, USA:

ACM Press, p. 570–575.

Sorgenfrei, C., K. Ebner, S. Smolnik and M. E. Jennex (2014)."From Acceptance to Outcome: To-

wards an Integrative Framework for Information Technology Adoption." In: Proceedings of the

Twenty Second European Conference of Information Systems. Tel Aviv.

Spagnolli, A., E. Guardigli, V. Orso, A. Varotto and L. Gamberini (2014). “Measuring User Ac-

ceptance of Wearable Symbiotic Devices: Validation Study Across Application Scenarios.” In:

Symbiotic Interaction. Ed. by G. Jacucci, L. Gamberini, J. Freeman and A. Spagnolli. Cham:

Springer International Publishing, p. 87–98.

Straub, D., M.-C. Boudreau and D. Gefen (2004). “Validation Guidelines for IS Positivist Research.”

Communication of the Association for Information Systems (13), 380–427.

Straub, D. W. (1989). “Validating Instruments in MIS Research.” MIS Quarterly 13 (2), 147–169.

Suh, B. and I. Han (2002). “Effect of Trust on Customer Acceptance of Internet Banking.” Electronic

Commerce Research and Applications 1, 247–263.

Swan, M. (2012). “Sensor Mania! The Internet of Things, Wearable Computing, Objective Metrics,

and the Quantified Self 2.0.” Journal of Sensor and Actuator Networks (1), 217–253.

Thompson, R. L., C. A. Higgins and Howell, Jane, M. (1991). “Personal Computing: Toward a Con-

ceptual Model of Utilization.” MIS Quarterly 15 (1), 125–143.

Thong, J. Y.L., S.-J. Hong and K. Y. Tam (2006). “The effects of post-adoption beliefs on the expec-

tation-confirmation model for information technology continuance.” International Journal of Hu-

man-Computer Studies 64 (9), 799–810.

Turel, O. (2016). “Untangling the complex role of guilt in rational decisions to discontinue the use of a

hedonic Information System.” European Journal of Information Systems 25 (5), 432–447.



Urbach, N. and F. Ahlemann (2010). “Structural Equation Modeling in Information Systems Research

Using Partial Least Squares.” Journal of Information Technology Theory and Application 11 (2),

5–40.

Venkatesh, V. and M. G. Morris (2000). “Why Don't Men Ever Stop to Ask for Directions? Gender,

Social Influence, and Their Role in Technology Acceptance and Usage Behavior.” MIS Quarterly

24 (1), 115–139.

Venkatesh, V., M. G. Morris, G. B. Davis and F. D. Davis (2003). “User Acceptance of Information

Technology: Toward a Unified View.” MIS Quarterly 27 (3), 425–478.

Venkatesh, V., Thong, James Y. L. and X. Xu (2012). “Consumer Acceptance and Use of Information

Technology: Extending the Unified Theory of Acceptance and Use of Technology.” MIS Quarter-

ly 36 (1), 157–178.

Wang, Y.-S., Y.-M. Wang, H.-H. Lin and T.-I. Tang (2003). “Determinants of User Acceptance of

Internet Banking: An Empirical Study.” International Journal of Service Industry Management 14

(5), 501–519.

Wold, H. (2004). “Partial Least Squares.” In: Encyclopedia of statistical sciences. Ed. by S. Kotz. Ho-

boken NJ: Wiley.

Yang, H., J. Yu, H. Zo and M. Choi (2016). “User acceptance of wearable devices: An extended per-

spective of perceived value.” Telematics and Informatics 33 (2), 256–269.

Yu, Y.‐T. and A. Dean (2001). “The Contribution of Emotional Satisfaction to Consumer Loyalty.”

International Journal of Service Industry Management 12 (3), 234–250.

INSIGHTS INTO PERSONAL ICT USE: … · Self-tracking, life-logging, quantified-self, personal...

Documents

Transcript of INSIGHTS INTO PERSONAL ICT USE: … · Self-tracking, life-logging, quantified-self, personal...